Cheap LLM Coding Models
Effective prices in US$ per 1B tokens, ordered by cache-hit cost. Lower is better.
Cache hit price
US$ per 1B cached input tokens
Subsidies already applied where indicated.
Price table
Search, sort, and change page size.
| # | Provider | Model | Condition | Cache hit / 1B | Cache miss / 1B | Output / 1B |
|---|---|---|---|---|---|---|
| 1 | OpenAI | GPT-5.6 Luna | ÷30 | $0.70 | $6.70 | $40.00 |
| 2 | Anthropic | Claude Haiku 4.5 | ÷100 | $1.00 | $10.00 | $50.00 |
| 3 | Anthropic | Claude Sonnet 5 | ÷100 | $2.00 | $20.00 | $100.00 |
| 4 | Muse Spark 1.3 Contributor | Contributor | $2.00 | $100.00 | $200.00 | |
| 5 | DeepSeek | V4.1 Flash | Off-peak | $3.00 | $150.00 | $600.00 |
| 6 | Anthropic | Claude Opus 5 | ÷100 | $5.00 | $50.00 | $250.00 |
| 7 | DeepSeek | V4.1 Flash | Peak-hour | $6.00 | $300.00 | $1,200.00 |
| 8 | OpenAI | GPT-5.6 Terra | ÷30 | $6.70 | $66.70 | $400.00 |
Removed from this comparison: Claude Fable 5.1, GPT-6 Astra, and GPT-5.6 Sol.
Values were converted from US$/1M tokens to US$/1B tokens (×1,000).